Texas A&M University  -  College of Architecture

 Course Home
 Syllabus
 Lectures/Labs
 Readings
 Assignments
 Students/Gallery

ARCH 653 Building Information Modeling in Architecture

Project 2 - Parametric BIM, Visual Programming, and AI Agents

Final version due: December 2, 2026 (Wednesday), 10:00 PM

Weight: 50%

Extend the functionality of parametric BIM using Revit and Dynamo (visual programming) to pass parameters, data, and calculations between different families and/or between a project and its families. Build on Project 1 by improving its models and adding new design ideas, data, content, or methods. You may also use another building design/model that is different from Project 1 with a comparable scope of work.

Generative AI may be used for conceptualization and visualization. You may also use AI to learn BIM methods, develop formulas, understand Dynamo nodes, and troubleshoot your work. Revit MCP and other AI/MCP tools may assist the modeling and visual programming process. The goal is to deepen your understanding of parametric BIM, data relationships, and related tools. You remain responsible for the submitted design, model, graph, and results.

Key learning outcome: With or without AI assistance, you must be able to explain your work, verify its accuracy and correctness, and present it clearly. A successful tool response or an attractive image alone does not demonstrate that the BIM or its logic is correct.

Project Requirements

1. BIM and Visual Programming Tutorials and Project Versions (15 pts)

Complete all homework/tutorials assigned during the second phase of the course and submit both checkpoint versions of this project. Completeness, timely submission, quality, and demonstrated progress count toward this part of the project grade. Follow the individual homework deadlines and submission procedures.

2. Parametric BIM through Visual Programming (45 pts)

Develop your own parametric BIM application using the methods studied in the course. Include at least one working Dynamo workflow that passes parameters or calculated results between multiple families and/or between the project and its families. Show how a change in an input produces the intended changes in the model. AI/MCP assistance is optional but encouraged to use for different components or steps of the modeling process.

Define driving parameters, their units and valid ranges, the relationships or equations they control, and the design purpose of those relationships. Organize and annotate the Dynamo graph so that you can explain its inputs, node groups, calculations, data flow, and outputs. Identify the parts you developed or modified from tutorials, samples, or AI-generated suggestions.

Possible applications include:

  • Use sun-path, weather, or other external data to drive window sizes, curtain panels, shading, or lighting-related design parameters.
  • Pass parameters among two or more families, such as floor dimensions controlling a roof and wall layout, or roof/wall angles controlling attached elements.
  • Use calculations, conditions, or iteration to explore a design rule or performance criterion. State assumptions and sources for any code- or performance-related claims.
  • Size or arrange hosted families or furniture according to room or host dimensions.
  • Use building functions or a shared project parameter to control multiple families or building configurations.
  • Create a parametric facade pattern or another creative application with clear design intent.
  • Use AI/MCP to query model data or assist a parameter-driven workflow, then verify the returned data and resulting changes in Revit and Dynamo.

Using AI / MCP to Enhance Learning

MCP (Model Context Protocol) connects AI applications with external tools and data. Depending on the capabilities of the installed connector, you may use Revit MCP or other AI/MCP tools to inspect model information, assist with creating or modifying BIM elements, set parameters, or help develop and debug Dynamo logic or scripts. Check what your specific tools support and document the Revit, Dynamo, AI, and MCP tools/versions you use.

Use the following AI process for a modeling or programming task:

  1. Plan and predict. Describe the design intent and expected result yourself. Identify the elements, parameters, units, constraints, and data involved before asking AI to help.
  2. Inspect and act in steps. Review the proposed operations, formulas, nodes, or scripts. Work on a saved copy and make changes in manageable steps so that you can inspect the effects.
  3. Verify and correct. Check the actual Revit model and Dynamo results against your prediction, measurements, calculations, or source data. Diagnose and correct errors, then rerun the checks.
  4. Explain and present. Describe what the tools did, why the method works, what you changed, and what evidence supports the result. Be prepared to modify an input and explain the resulting behavior.

For example, you might ask an AI/MCP tool to inspect room areas and relevant family parameters, propose a Dynamo relationship between room width and furniture size or location, and help implement it. You would then check the selected rooms, parameter names and units, calculated sizes/location, and actual family geometry yourself, including behavior at a limiting room size.

Verification evidence: Test at least three distinct parameter sets. Record the inputs, expected results, actual results, and pass/fail conclusions in a compact table, supported by screenshots, schedules, measurements, or independent calculations. Where relevant, also test an invalid input and explain how it is handled. Check parameter types and units, family/project relationships, geometry, and Dynamo warnings or errors. For AI/MCP actions, confirm that the intended elements changed and check for unintended changes or duplicate elements when the workflow is rerun.

AI/MCP disclosure: If you use these tools, include representative prompts/requests, important tool actions or generated logic, what you accepted or changed, errors you found and corrected, and how you verified the final result. State which parts were developed directly by you and which were assisted by AI. If you do not use AI/MCP, state that briefly. Label AI-generated conceptual images and visualizations and explain any differences from the modeled geometry or data.

3. Visualization and Documentation (20 pts)

Publish the project documentation on your portfolio/blog, or submit a PDF through Canvas as described below. Include:

  • The building/application background, design intent, and improvements or new work beyond Project 1.
  • Diagrams and equations explaining the parameter relationships and workflow, with clear images of the annotated Dynamo graph and resulting BIM.
  • Screenshots showing different parameter states, relevant Revit views, and data displays.
  • The verification table and supporting evidence, corrected errors, remaining limitations, and instructions for reproducing the work.
  • The AI/MCP disclosure, if applicable, and references for all external data, code, graphs, images, and other sources. If you use AI visualization, identify it and discuss its relationship to the verified BIM.
  • A brief reflection on what you learned about parametric BIM and visual programming, and how AI/MCP affected your understanding, effort, or results if used.

Create a narrated screen-recording movie of up to 4 minutes (strict). Explain the design and parameter relationships, show the Dynamo workflow and parameter changes in Revit, summarize verification evidence, and explain your AI/MCP contribution if applicable. Use your own voice to demonstrate your understanding. Suggested resolution: 1920x1080 pixels; suggested format: MP4. Clear images and sound are expected.

Include the movie in your Canvas submission. You may also embed it in your portfolio/blog directly or through YouTube or Vimeo. See Blogger Help: Add images and videos to your blog.

Give a reference for each image or citation of other people's work. Follow the AI Usage Policy: Guidelines and Disclosure in the course Syllabus (which links to Simple Syllabus).

4. Final Presentation (10 pts)

Present your project movie and participate in Q&A (if time permits). Be ready to explain the model structure, parameter relationships, Dynamo logic, verification methods, and the role of AI/MCP in your process. You should be able to explain or reproduce a key step and interpret the result of a parameter change. Make sure the movie and supporting files are accessible before the presentation.

5. Submission

Submit the following items by December 2, 2026, 10:00 PM.

5.1 Project Files to Canvas

Project folder name Project2_Final_YourFirstname_YourLastname
Files in the folder Organize your Revit project and required family files (.rvt/.rfa), Dynamo graphs (.dyn), supporting scripts and input data, images/screenshots, verification evidence, AI/MCP disclosure if applicable, documentation, and presentation movie into clear subfolders.
File to submit Project2_Final_YourFirstname_YourLastname.zip
Submit to Canvas
ARCH653 > Project 2 Final

5.2 Portfolio/Blog or PDF Documentation for Presentation

Create a portfolio/blog page containing your documentation, images, and presentation movie. Include your name and ARCH653 Project 2. Submit the URL to Canvas when required. As in Project 1, the class presentation links will be collected in an access-controlled spreadsheet hosted on TAMU Google Drive. You may control access to your portfolio/blog; arrange instructor access before the presentation if it is restricted.

If you do not use a blog or online portfolio, submit PDF documentation to Canvas and include the movie as a separate file. A link to the movie may also be included in the PDF. Make sure the instructor can access all presentation materials.

The Students/Gallery page will be used for presentation.

5.3 Project Thumbnail to Canvas

Create a 160x120 pixel landscape thumbnail from a representative view or rendering of your project. Save it as Project2_YourFirstname_YourLastname.jpg and upload it to the Project 2 Thumbnail assignment in Canvas. Check your submitted files, documentation, movie playback, and any shared links before the deadline.

Learning Objectives

  • Develop and explain parametric BIM relationships across Revit families and projects.
  • Use visual programming, data, and calculations to support a clear design intent.
  • Use AI/MCP, when chosen, to enhance learning and implementation while retaining responsibility for the work.
  • Independently verify model and computational accuracy, identify errors and limitations, and correct the work.
  • Present and defend the design, methods, results, and verification evidence clearly.

Evaluation

Your work will be judged on its workload, richness of design thinking, technical quality, visual quality, documentation, and demonstrated understanding. The point distribution remains: tutorials/checkpoints 15, parametric BIM and visual programming 45, documentation 20, and presentation 10. Verification supports the technical assessment; your explanations, disclosure, and presentation demonstrate your understanding.

About Exceptional Work (10 pts)

As stated in the Syllabus, 90/100 is the basis for satisfactory work that meets the standards. Errors reduce points; exceptional work may earn up to 10 additional points. Examples include a sophisticated, well-integrated parametric application with strong design intent and excellent verification; a well-supported simulation or computation; or creative, responsible AI/MCP use that demonstrates deeper understanding, clear documentation, and independently verified results.

How to Get Started and Resources

Review Homework 7 and 8 and the course lectures/labs. Study a relevant sample, explain how it works, and adapt it to your design intent. Check whether the sample runs in the Conceptual Mass Family editor or in a Revit project, and use files compatible with your software version.

  • RoomTableChair sample archive: a course example of using Dynamo for a parametric project. Review the included versions and update dependencies as needed.
  • Dynamo Primer: background and tutorials on visual programming, data, geometry, and Revit workflows.
  • Official MCP introduction: background on connecting AI applications to tools and data. Use the documentation for your installed Revit/AI connector for specific capabilities and setup.
  • AI Rendering for Architects (Revit + Veras): an optional visualization reference. Distinguish generated imagery from verified model geometry.